[1] 11w Changes in Ca channel expression upon differentiation of SN56 cholinergic cells

[1] 222w The SN56 cell line, a fusion of septal neurons and neuroblastoma cells, has been used as a model for central cholinergic neurons. These cells show increased expression of cholinergic neurochemical features upon differentiation, but little is known about how differentiation 21 affects their electrophysiological properties. We examined the changes in Ca channel expression that occur as these cells undergo morphological differentiation in response to serum withdrawal and exposure to dibutyryl-cAMP. Undifferentiated cells expressed a T-type current with biophysical and pharmacological properties similar, although not identical, to those reported for the current generated by the 21 a (CaV3.2) Ca channel subunit. Differentiated cells expressed, in addition to this T-type current, high voltage activated currents 1H which were inhibited 38% by the L-type channel antagonist nifedipine (5 mM), 37% by the N-type channel antagonist v-conotoxin-GVIA (1 mM), and 15% by the P/ Q-type channel antagonist v-agatoxin-IVA (200 nM). Current resistant to these inhibitors accounted for 15% of the high voltage activated current in differentiated SN56 cells. Our data demonstrate that differentiation increases the expression of 21 neuronal type voltage gated Ca channels in this cell line, and that the channels expressed are comparable to those reported for native basal forebrain cholinergic neurons. This cell line should thus provide a useful model system to study the relationship between calcium currents and cholinergic function and dysfunction.

[1] 45w SN56 cell line, formed as a fusion between neuroblastoma N18TG2 cells and mouse septal neurons, expresses spe-Dysfunction of central cholinergic neurons has been cific cholinergic markers including choline acetyltransimplicated in age-related neurodegeneration and in several ferase, acetylcholinesterase, and the vesicular acetycholine diseases including Alzheimer's [6,9,12,13,31,37,43,66].

[2] 529w transporter, and has been increasingly used as a model to Cell lines which faithfully reproduce the cholinergic study cholinergic neuron function and dysfunction phenotype may be useful to investigate the physiological [7,10,14,16,34,36,45,46]. and pathological alterations present in such states. The Despite the evidence for neurochemical similarity between the SN56 cell line and native cholinergic neurons, there is conflicting data as to whether these cells express the Ca channels expected for the cholinergic phenotype. HVA, High voltage activated; LVA, Low voltage activated; MB1, Previous electrophysiological studies failed to detect bio-Mibefradil 1; NMDG, N-methyl-D-glucamine; Nif, Nifedipine; sAHP, physically or pharmacologically defined neuronal type Slow afterhyperpolarization 21 High Voltage Activated (HVA) Ca currents [14]. How-*Corresponding author. Tel.: 155-31-3499-2707; fax: 155-31-3499-21 ever Ca imaging experiments have shown a component 2695. 21 1 E-mail address: chris@icb.ufmg.br (C. Kushmerick). of Ca increase induced by K depolarization in con- fluent SN56 cells that is blocked by the N-type Ca upright microscope with a 403 water-immersion phase channel blocker, v-conotoxin-GVIA [16]. Furthermore, contrast objective. Capacitive currents were electronically differentiated SN56 cells express the a (Ca 2.1; nomen-compensated and a P/ 4 protocol [1] was used for linear 1A V 21 (leak and residual capacitive) current subtraction. The clature as in Ref. [18]) voltage gated Ca channel subunit standard extracellular solution contained (in mM), NaCl responsible for P/ Q type currents which is sensitive to the 21 (125), BaCl (25), HEPES acid (10), CsCl (5.4), glucose specific blocker v-Agatoxin-IVA, as determined by Ca 2 (10), pH57.4 (adjusted with NaOH). In control experiimaging [7]. 21 ments performed as previously described (see Ref. [32]), The types of Ca channels that mediate ACh secretion 1 the Na current present in both undifferentiated and vary greatly between cholinergic cell types. In many differentiated SN56 cells was inhibited .95% by tetcholinergic cell types, from the electrocyte in electric fish rodotoxin (TTX) with an apparent dissociation constant of [54,57, but see 53] to mammalian central neurons [60,63] |5 nM. Accordingly, when solutions containing Na were or peripheral neurons [15,21,29,49,67, but see 58], L-or 21 used to measure Ba currents, TTX was added to the N-or P/ Q-type channels, or a combination of the three are extracellular solution at 1 mM. On some occasions (|13% necessary for ACh secretion. Of particular relevance to the 21 of experiments) Ba currents recorded in the presence of SN56 cell line is the observation that acutely isolated basal 1 mM TTX appeared to contain small amounts of TTXforebrain cholinergic neurons from rat [2,4] and guinea pig 1 resistant Na currents. This current, which occurred in [69] express both Low Voltage Activated (LVA) T-type as 21 both undifferentiated and differentiated cells, was almost well as L-, N-, P/ Q-and R-type HVA Ca currents, the 21 21 always much smaller than the peak Ba current, and had latter three being important for Ca mediated exocytosis much faster inactivation kinetics, and was thus readily of ACh and its inhibition by autoreceptors [2,3]. 21 separable from the HVA currents described herein. None-Ca has a pivotal role in a number of cell processes, theless, since some of the inhibitors of LVA channels used from development to cell death, and the cellular responses 21 1

[3] 47w herein (Ni , mibefradil) are also active on Na channels, induced by this second messenger depends on its entrance the pharmacological characterization of the LVA current pathway [42,69]. Therefore, the use of SN56 cells as a with CsOH. For studies in Na -free solutions, Na ATP 2

[1] 286w withdrawing the sera from the media and elevating in-Differentiated cells expressed Ca channel currents that tracellular cAMP levels with dibutyryl-cAMP. This treatwere larger, and differentiation was correlated with the ment causes morphological and neurochemical changes in appearance of a sustained component of current. Differenthese cells, as previously described in detail [7,10]. tiated cells had peak and sustained current densities of 21 21 Undifferentiated cells were spheroidal or bulbous, with-16.560.80 pA pF and 8.160.61 pA pF , respectively. out any axon-or dendron-like processes and had mem-As will be shown below, the currents obtained from brane capacitance of 3062.6 pF (n547). Cells that were undifferentiated cells had pharmacological and biophysical 21 Fig. 1. Changes in Ca channel currents upon differentiation in SN56 21 21 21 Ni exhibited by the LVA current in undifferentiated cells. Whole-cell Ca channel currents (25 mM Ba as the charge SN56 cells is similar to that seen in the currents generated carrier) in SN56 cells, recorded during voltage steps to 0 mV from a holding potential of 280 mV. Cells were plated onto glass coverslips and by the the a (Ca 3.2) T-type Ca channel [35,47]. This 1H V maintained for 2-4 days either in normal culture medium (undifferenchannel is reported to have a higher permeability for Ba tiated, top) or induced to undergo differentiation as described in the compared to Ca [38]. To characterize the divalent ion Methods section (bottom). The thick black lines are the mean current selectivity of the LVA current in undifferentiated SN56 densities obtained from different cells (undifferentiated, n520 cells; cells, we compared the current measured using 25 mM differentiated, n580 cells). The thin black lines are the 5th and 95th 21 21 percentile currents (ranked by peak amplitude).

[2] 58w Ca or 25 mM Ba as the charge carrier (Fig. 2C). When measured at 0 mV or 130 mV (to avoid any effects on gating and to observe pure effects on permeation) 21 21 replacing Ba with Ca did not cause a statistically significant change in the peak current (reduction of 11614% and 13620%, mean695% confidence limit, respectively).

[3] 90w Other characteristics of LVA channels are their high sensitivity to the drug mibefradil 1 (MB1) and low 21 sensitivity to blockade by Cd . We tested MB1 at 1 and 10 mM and found that it inhibited 5963% and 9262.3% 21 respectively of the Ca channel current in undifferentiated SN56 cells (Fig. 2D). This current was also much 21 21 more sensitive to inhibition by Ni than Cd ; whereas 21 16 mM Ni blocked 4462% of the current, the same 21 concentration of Cd blocked only 1264% (Fig. 2D).

[4] 234w As shown in Fig. 1, Ca channel currents obtained from cells that had been induced to differentiate exhibited a sustained current that did not inactivate during the 200 ms voltage pulse. This observation was further explored using a standard two-pulse protocol in which the current available in the second pulse is measured as a function of the potential during a pre-pulse (Fig. 3). Stepping undifferentiated SN56 cells to depolarized potentials for 200 activation near 243 mV) and a sustained component SN56 cells, obtained during voltage pulses to 220 mV (top) or 0 mV (about 50% of the total current), which remained available (bottom) following 200 ms pre-pulses from 260 mV to 225 mV (top) or even after pre-pulses to potentials as high as 120 mV. from 260 mV to 215 mV (bottom) in 5 mV increments. (B) Averaged h -V relationships for undifferentiated (circles, n57 cells) and differentiated (squares, n56 cells) SN56 cells. Abscissas is the voltage during the 200 ms prepulse. the smooth lines represent the best-fit to Eq. (4), a 3.4. Activation properties of differentiated cells single Boltzmann inactivation process plus a constant (non-inactivating) term. Parameters thus obtained for undifferentiated cells were, V 5 1 / 22inac Differentiated cells expressed, in addition to the LVA 245 mV, k 56.9 mV, 95% inactivation. Parameters for differentiated inact component described above, sustained currents which cells were, V 5243 mV, k 58.7 mV, 58% inactivation.

[5] 71w 1 / 22inact inact activated at potentials more positive than 210 mV and had peak current at potentials more positive than 0 mV (Fig. 4). The averaged I-V relationship for the peak inward current from differentiated cells consisted of a single broad distinct shoulder in the I-V and g-V relationships (Fig. 4C asymmetric peak centered at approximately 0 mV which and D). could be modeled as two populations of channels, HVA

[6] 182w The LVA component in both differentiated and unand LVA (Fig. 4B). Although well determined by the differentiated cells were compared at 230 mV (to avoid curve-fitting procedure used here, the overlaps of the two activating HVA channels in the differentiated cells). Meacomponents makes it difficult to see the two distinct sured at this potential, the LVA current exhibited a small 21 populations in the raw data. In order to improve the but statistically significant decrease in Ni sensitivity separation, we constructed the I-V relationship for the (Fig. 2D) and a decrease in the rate constant of inactivacurrent measured 50 ms after the onset of the voltage tion (4863.7 ms n510 undifferentiated cells versus pulse, at which point the LVA component had partially 3762.2 ms n530 differentiated cells, P,0.02). Thus, the decayed. Under these conditions, one can easily discrimi-LVA current undergoes some subtle changes during difnate the two components, as the LVA current makes a ferentiation in SN56 cells. 3.5. Pharmacology of the HVA component the currents shown were obtained by normalizing and averaging over several cells (see the Methods section and 21

[7] 237w In order to determine the types of Ca channels Figure legends for details). We also tested 5 mM (6)-Bay responsible for the HVA component of the current in K 8644, a specific activator of L-type channels, and found differentiated SN56 cells, we studied the effect of drugs that it increased the sustained component of current in and toxins with known specificities on this current. differentiated SN56 cells by 52610% (n55 cells; not Nifedipine (5 mM), one of the more specific and potent of shown). Taken together, these observations imply that a the 1,4-dihydropyridine (DHP) L-type channel blockers major component of the HVA current present in differen-[20], reversibly inhibited the sustained current by 3865% tiated SN56 cells is carried by L-type channels. (n57 cells). Under identical conditions, 5 mM nifedipine The rapidly inactivating LVA current was also sensitive blocked 7065% of the L-type current present in GH3 cells to nifedipine, but its affinity was much lower. In order to (n55 cells, not shown). As shown in Fig. 5A the current obtain 50% inhibition, a concentration of about 30 mM inhibited by nifedipine (obtained as the difference between was necessary (not shown). Inhibition of neuronal T-type the control current and the current resistant to the drug) channels by relatively high concentrations of nifedipine exhibited little or no inactivation, as expected for L-type and other DHPs has been reported previously [5,52,59]. currents measured with EGTA in the pipette solution and

[8] 209w In order to determine which other types of neuronal 21 using Ba as the charge carrier [26]. Note that in Fig. 5, HVA channels are expressed by differentiated SN56 cells, we used two specific toxins, v-CTx-GVIA and v-Agaof application [50]. The second toxin, v-Aga-IVA, inhibits 21 IVA. The first of these is a specific and irreversible P and Q type Ca channels with IC values on the order 50 inhibitor of N type channels [44]. The concentration we of 1 and 90 nM, respectively [39,50]. Recently, it has been used (1 mM) is capable of completely inhibiting the N-type shown that at concentrations approaching 1 mM, v-Agachannels present in cerebellum granular cells within 100 s IVA also inhibits N type channels [56]. To avoid this possibility, we used the toxin at a concentration of 200 differing observations cited above, we sought to character-21 nM. We, thus, expected to block .99% of P type channels ize Ca channel expression in SN56 cells and to charand |70% of Q type channels present, without sacrificing acterize the changes that occur during differentiation. specificity. The two toxins tested did not affect the LVA 21 component of the Ca channel current in differentiated 4.1. The LVA component cells, measured at 230 mV (not shown).

[9] 242w As shown in Fig. 5B, 1 mM v-CTx-GVIA rapidly LVA T-type currents are widely expressed in the central inhibits a component of the sustained current obtained at 0 nervous system and play important roles in the regulation mV in differentiated cells. The toxin-sensitive component, of neuronal behavior [28]. Until recently, the lack of obtained by subtraction as described above, exhibits a knowledge of the molecules responsible for these currents, somewhat slower activation than the nifedipine sensitive together with their poorly defined pharmacology prevented current, with little or no inactivation during the voltage the degree of characterization of T-type channels that had 21 pulse. After 150 s treatment with toxin, the current had been achieved with L-, N-, P/ Q-and R-type HVA Ca decreased by 4463%. If we assume that the run-down channels. With the recent cloning of LVA channel isoforms during these sets of experiments was similar to that which together with advances in our understanding of their occurred during the (identically performed) experiments pharmacology, this situation is rapidly changing and it is with nifedipine (|5% per 100 s), then this is an overesti-now possible to identify the distribution of T-type channel mate of the fraction of N-type current in the HVA isoforms in the brain and, based on their differing biocomponent, and the real value would be 3764%. physical properties, postulate specific functions that each The current inhibited by 200 nM v-Aga-IVA can be subserve [33,35,38,40,41,47,69]. distinguished from the nifedipine-and v-CTx-GVIA-sensi-

[10] 306w In basal forebrain cholinergic neurons, one role of Ttive current based on its relatively fast activation and the type currents may be to control firing frequency. These fact that it inactivated by 50% during the course of the cells express an apamin-sensitive slow afterhyperpolarizavoltage pulse with a time constant of 14 ms (Fig. 5C). As tion (sAHP) which is coupled to firing frequency [23]. In reported by others [50,56], inhibition by v-Aga-IVA was guinea pig Nucleus Basalis cholinergic neurons, this sAHP 21 also slower than that by nifedipine and v-CTx-GVIA. is activated by Ca influx through T-type channels based After taking into account run-down, we estimate the on its inhibition by amiloride [69]. Similarly, in rat Medial 21 contribution of P/ Q type channels to the HVA Ca Septum and Nucleus of the Diagonal Band of Broca channel current in these cells to be 1565%. neurons, activation of the sAHP is blocked by low 21 21 The effects of the three Ca channel antagonists used concentrations of Ni [22], suggesting activation via Tin the experiments described above were additive. A type channels, although R-type channels could also be combination of nifedipine (25 mM), v-CTx-GVIA (1 mM) involved. and v-Aga-IVA (200 nM) inhibited 8561% (n52; not SN56 had well developed LVA calcium channel curshown) of the HVA current present at 0 mV. This is rents. This current is distinguishable from a1 R-type E somewhat less than the sum of the inhibitions observed currents because the latter has much slower inactivation when the drugs and toxins were applied separately, imply-kinetics [30,62] lacks the criss-crossing pattern described ing that there is some overlap in their specificity. The by Randall and Tsien [51], and has a much higher 21 21 current resistant to the combination of nifedipine, v-CTxsensitivity to Cd (the IC for inhibition of a1 by Cd 50 E

[11] 20w GVIA and v-Aga-IVA was largely inhibited by a combina-is 4 mM; Ref. [30]). The LVA current in SN56 cells has

[12] 209w 21 21 21 tion of 200 mM Ni and 200 mM Cd , indicative of gating kinetics and voltage dependence and a Ni sen-21 current through Ca channels. Thus about 15% of the sitivity similar to those reported for the LVA currents from current can be classified as R type. basal forebrain cholinergic neurons [4]. Three isoforms of T-type channel a1 subunits have been cloned thus far, a (CaV3.1), a (CaV3.2), and a 1G 1H 1I 4. Discussion (CaV3.3), with differing pharmacological and biophysical properties [38,40,41,68]. The LVA current in SN56 cells 21 The use of appropriate cells lines can help in understand had kinetic properties and Ni sensitivity most similar to aspects of cholinergic function and dysfunction. Despite those for a [8,35,38]. Moreover both the LVA currents in 1H the evidence that SN56 cells, upon differentiation, correct-SN56 cells and current through a have an IC50 for 1H ly express several biochemical markers for the cholinergic inhibition by mibefradil of |1 mM and show a biphasic 21 phenotype, there is some disagreement as to whether they dose-response curve for inhibition by Ni [47]. On the 21 express neuronal Ca channels similar to native forebrain other hand, the a pore has been reported to be much 1H 21 21 21

[13] 133w cholinergic cells [7,14,16]. Ca has multiple roles in cell more permeable to Ba than Ca [38] whereas for the survival and function, and thus whether SN56 cells will LVA currents in SN56 cells the two ions had comparable provide an adequate cholinergic model may depend on the permeabilities. Thus, the LVA current in SN56 cells is 21 expression of appropriate Ca channels. In view of the similar but not identical to those reported for a . Species 1H specific properties, alternative splicing and technical issues of the L-type current in GH3 cells as an internal control. such as choice of divalent ion concentration may account Our estimate of the amount of L-type current in differenfor the observed differences. tiated SN56 cells falls within the range reported for native basal forebrain neurons (Table 1).

[14] 107w 38%, 37%, and 15%. The R-type current, measured during However our data agree with previous results using Ca 21 application of nifedipine plus the toxins was 15%. fluorimetry which indicate that neuronal type Ca chan-Pharmacological dissection of the HVA currents has nels are present in SN56 cells [7,16]. been carried out in basal forebrain neurons from rat [2] and guinea pig [24,69] permitting comparison with our results 21 4.3. Changes in Ca channel expression during (Table 1). N-type currents have been estimated to make up differentiation 35-45% of the HVA current in native cells, similar to what we observed in SN56 cells. The fraction of HVA current

[15] 109w The morphological and neurochemical differentiation that is dyhydropyridine-sensitive (i.e. L-type) reported for that occurs in SN56 cells during serum withdrawal and native cells ranged from 10% to 50%. Part of this incubation with dibutyryl-cAMP (8-bromo-cAMP is also difference may arise due to different concentrations of effective, unpublished observations) are accompanied by nifedipine used (3-10 mM) to determine the size of this 21 component. We opted for 5 mM nifedipine as a com-changes in the expression of Ca channels. Prior to promise between efficacy and specificity, using inhibition differentiation, SN56 cells expressed exclusively LVA 21 Ca channels whereas differentiation turned on expression of HVA channels found in native cholinergic neurons.

[1] 15w was omitted from the pipette solution and was substituted by an additional 7 mM TEA-Cl.

[2] 54w 2.1. Cell culture SN56 cells were a gift from Professor Bruce Wainer 2.2.2. Electrodes (Department of Pathology, Emory University School of A Narishige PP83 vertical pipette puller was used to Medicine). The cells were maintained in DMEM (Sigma) make electrodes from GC150F borosilicate glass (Clark). supplemented with 10% fetal bovine serum (Gibco), 2 mM

[3] 45w When filled with pipette solution, whole-cell electrodes L-glutamine (Sigma) and 1% penicillin / streptomycin, in a had resistances ranging from 1.5 to 3.5 MV. The patch 5% CO atmosphere at 378C. Cells were plated onto glass clamp amplifier ground was connected to a small side

[4] 85w coverslips at a density of about 10 cm and maintained chamber containing 3 M KCl with a Ag-AgCl pellet, and either in the culture medium described above (undifferenthe side chamber was connected to the recording chamber tiated cells) or induced to differentiate by withholding the via a glass bridge filled with 4% agar in 150 mM NaCl. serum from the medium and including 1 mM dibutyryl-cAMP. Medium was changed every 48 h except during 2.2.3. Voltage control differentiation when it was changed every 24 h.

[5] 14w Series resistance was compensated as necessary to maintain the voltage error,5 mV (typically,2 mV).

[6] 138w Differentiated cells occasionally exhibited signs of poor voltage control such as, (1) slow or asynchronous (with 21 2.2.1. Ca channel current respect to the voltage pulse) activation, (2) slow or An Axon 200B patch-clamp amplifier was used at room asynchronous or incomplete deactivation, (3) membrane temperature (20-258C) to measure whole-cell ion currents capacitance that could not be almost completely compen- [25]. Patch clamp experiments were performed in a small sated by the single R-C compensation circuit of the patch (400 ml) glass-bottomed chamber using an Zeiss Axioskop clamp amplifier. Data from cells that exhibited one or more 21 of these symptoms were excluded. Data are not corrected ment at |1 ml min . Solutions were gravity fed to the input ports of a six-way solenoid valve (Cole-Parmer) for the pipette junction potential of |7 mV (i.e. V 5

[7] 37w Membrane mounted close to the recording chamber. A thin plastic V 27 mV), as calculated using the junction potential Pipette tube was connected to the output of the valve and attached tool in pClamp 8.0 (Axon Instruments).

[8] 111w to the microscope objective such that the desired solution was ejected into the microscope field of view (and thus 2.2.4. Data acquisition and analysis onto the cell under study). This arrangement permitted Voltage commands and acquisition of data from the rapid solution changes which was confirmed in each patch clamp amplifier were achieved using a DigiData 1 experiment by observing the disappearance of the Na 1200 digitizer controlled by pClamp 8.0 software (Axon current when TTX was added. When peptide toxins were Instruments) connected to a PC computer (Dell). Currents used, BSA (0.05%; Sigma) was included in all solutions to were low-pass filtered at 5 kHz and sampled at 11.1 kHz.

[9] 148w avoid non-specific binding. Mibefradil (Merck) was stored Curve-fitting was performed using SigmaPlot (SPSS, Inc). at 100 mM in water and diluted to the final concentrations Cell membrane capacitances were measured by reading the directly in the extracellular solution. Nifedipine (Sigma) value of the slow capacitance compensation control on the and (6)-Bay K 8644 (RBI) were made up as a 10003 patch clamp amplifier after canceling the capacitive transstock solution in DMSO. In experiments with these drugs, ients. This value was confirmed using the Membrane Test all solutions contained 0.1% DMSO, which had no detectprogram built-in to the pClamp software (with capacitance 21 able effect on the Ca channel currents. Nifedipine was compensation turned off). This program analyzes the used at a concentration of 5 mM except when used in response to a square voltage step to estimate the whole-cell combination with the peptide toxins to measure R-type parameters.

[10] 41w currents; in these cases it was used at 25 mM to obtain Current-voltage (I-V ) data were used to adjust an maximal inhibition and to overcome binding by BSA. The equation of the form: peptide toxins v-CTx-GVIA and v-Aga-IVA were from

[11] 50w Peptides International, and were stored as 100-5003 stock solutions in deionized water. When these toxins were (1) used, an equivalent amount of deionized water (0.2-1%) to determine the maximum conductance, g , the effecmax was added to all solutions. tive reversal potential, V , the midpoint of activation rev (V

[12] 10w ) and the steepness factor (k ). When two com-

[13] 50w 1 / 22act act 2.3. Statistics ponents of current were present (LVA and HVA) the equation used to fit the data consisted of the sum of two Unless otherwise stated, mean values are given as such equations, with separate values of g , V , and max, 1 / 22act

[14] 34w mean61 SEM. The confidence intervals for the mean k for each component. From these data, conductanceact current traces shown in Fig. 5 were obtained by calculating voltage ( g-V ) relationships were determined as:

[15] 19w the standard error over all current traces at each time point and then calculating the 95% confidence limit as:

[16] 20w rev 95% Conf 5 t (N 2 1,0.95) ? SEM and used to adjust g-V curves based on the equation:

[17] 35w max 1 / 22act act non-parametric statistical procedure (resampling) was also or the sum of two such equations when both LVA and HVA used to determine the confidence intervals about the mean current were present.

[18] 24w current. The results obtained using resampling agreed Steady-state inactivation (h ) was determined from the within,5% with the results obtained using the t statistic.

[19] 27w current measured during voltage pulses preceded by 200 Currents corresponding to the 5th and 95th percentile were ms pre-pulses to various potentials from 2100 to 20 mV.

[20] 62w chosen by sorting the currents by peak amplitude and then The amplitude of the peak inward currents were normalchoosing the current closest to the desired ranking. Comized to that following the pre-pulse to 2100 mV and used parison between mean values was done using the Student's to adjust an equation of the form: t-test. Unless otherwise specified, the unpaired t-test was used.

[21] 22w where a is the fraction of current that did not undergo inactivation, and V and k are the mid-point and 3. Results

[22] 12w the steepness factor for the fraction of current susceptible 21 to inactivation.

[23] 34w Differentiation of SN56 cells by dibutryl-cAMP has been shown to enhance evoked release of ACh from these The treatment we employed to induce differentiation cells [10], as well as the expression of synaptic vesicle

[24] 88w 21 causes SN56 cells to change from round cells apparently protein Synaptotagmin I [7]. The changes in Ca channel 21 expressing exclusively LVA Ca channels to morphologi-expression which we describe may account for part of the 21 cally differentiated cells expressing the HVA Ca channel enhancement in ACh release. In guinea pig basal forebrain types found in native cholinergic neurons. Furthermore, neurons, liberation of ACh during evoked action potentials 21 this transition is complete within 2 days, and cells that required simultaneous Ca influx through both N-and 21

[25] 39w were not subjected to this treatment continue to grow, but Q-type Ca channels (R-type channels acting alone could do not express significant HVA currents even after 4 days also support release, but at much lower levels, whereas Lin culture.

[26] 9w and T-type channels were unable to support release) [2].

[27] 196w 21 Nifedipine (5 mM) inhibited 38% of the HVA current in The change in Ca channel expression in SN56 cells, differentiated SN56 cells. Thus it seems clear that under from exclusively T-type to a mixture that includes N and the conditions we used, a large percentage of the HVA P/ Q-type would thus be expected to provide these cells 21 21 Ca current is L-type. Our results with (6)-Bay K 8644 with the Ca channel complement necessary to impleconfirm this conclusion. Inhibition with the toxins v-CTxment the cholinergic phenotype. GVIA and v-Aga-IVA produced a clear blockade of Colom et al. [14] surveyed the ionic currents in di-21 another large fraction of the HVA Ca channel current, butyryl-cAMP differentiated SN56 cells using voltage and this inhibition was additive with that induced by clamp protocols similar to our own, and did not detect the nifedipine. Based upon the percent inhibition by each of v-toxinand nifedipine-sensitive HVA type currents we these inhibitors applied separately we estimate the contri-report. We do not know why the HVA currents that we bution of L-, N-, and P/ Q-type to the HVA current to be describe were not observed in these previous experiments.

[1] 67w The bath was superfused continuously during the experi-Whole cell patch clamp experiments were carried out on 1 SN56 cells plated for 2-4 days and either maintained in the experiments were performed with Na -free solutions complete culture media (herein referred to as undifferen-(see the Methods section), eliminating the possibility that 1 tiated cells), or induced to undergo differentiation by these currents are carried by TTX-resistant Na channels.